Research On Industrial Robot Assemble System Based On Stereo Vision

Social productivity has been highly promoted with the rapidly development of the scienceand technology. And so the robot has been born so as to meet the develop need. Nowadays,industrial assembly robot was occupy an increasingly important position in various industrialfields. Since the industrial robot has experienced three period, online teaching, offlineprogramming and robot added vision. Robot can reconstruct3D information, intelligentreactive with the environment after it added vision. But due to the need for processingmulti-media such as image, the computing amount was greatly increased, thus lead to a lowreal-time system performance.Visual attention mechanism select the specific outstanding area from the visual inputinformation based on the principle of bionics. The workpiece search speed can be highlyaccelerate by creatively add visual attention mechanism into traditional visual system, thusimprove the entire system computing performance. Visual system can intelligent reconstructthe3D information by add visual attention mechanism, intelligentialize the system at utmost.Research about robot at home and abroad starts from1970s, from online teaching,offline programming to vision added, robot intelligent level was higher and higher. But theresearch about add visual attention mechanism into traditional visual system is not so much.The robot has to process more and more information, and the computing amount was biggerand bigger, the limited computing resource restrict the development of robot. This paper isunder such a actual background, creatively add visual attention mechanism into visual systemand do some experiment based on the existing industrial robot platform in my laboratory. Theexperiment include saliency map obtain, camera calibration, workpiece extraction andrecognition, homogeneous frame transformation, inverse solution generation, driver robot.through a series experiment, reappear the reality usage scene of robot visual system, andcompare this method with the traditional visual system, the result indicates that the methodadded visual attention mechanism has obviously advantages.